Enterprises are deploying autonomous AI agents faster than any security framework was designed to handle them. Trent AI, a London-based agentic security company, is building the infrastructure answer — and it raised $13 million in seed funding to do it. The company emerged from stealth on 7 April 2026 with a multi-agent security platform purpose-built for the agentic security multi-agent enterprise environment, backed by LocalGlobe and Cambridge Innovation Capital. Read about multi-agent security platform, Trent AI.
The problem Trent AI is solving is structural, not theoretical. According to Deloitte, 74% of enterprises plan to deploy agentic AI within two years. Only 21% report a mature governance model for autonomous agents. That gap does not close on its own. It widens with every production deployment that ships without a dedicated security layer.
Conventional security tools — SIEMs, CSPMs, static rule engines — were designed for environments where humans make decisions and systems wait for approval. Agentic AI operates on a different clock. Agents scan infrastructure, call APIs, write and execute code, and modify configuration at machine speed. A security framework that requires human review at each step is not a control — it is a bottleneck.
How the Platform Works
Trent AI deploys four specialised agent types running continuously in parallel. Scan agents monitor code, infrastructure, dependencies, and runtime behaviour to locate risk as it emerges. Judge agents classify and prioritise signals based on actual business impact rather than predefined rule sets. Mitigate agents patch vulnerabilities and validate fixes automatically without waiting for a human approval queue. Evaluate agents track risk trends over time and benchmark the deployment against established standards.
The feedback loop between these four layers is the core architectural decision — each cycle improves the accuracy of the next. The system is not a scanner that reports; it is a runtime security layer that acts.
[INTERNAL LINK: AAI article on multi-agent orchestration enterprise architecture]
Founding Team and Credibility Signal
The founding team is calibrated for enterprise trust. CEO Eno Thereska was a Distinguished Engineer at Alcion (acquired by Veeam), AWS, and Confluent. Co-founder Neil Lawrence holds the DeepMind Professorship of Machine Learning at the University of Cambridge and served as Amazon’s Director of Machine Learning. Co-founder Zhenwen Dai was a machine learning scientist at AWS and Senior Research Manager at Spotify.
Angel investors include Joaquin Quinonero Candela (OpenAI), Avinash Bhat (AWS), Ippokratis Pandis (Databricks), and Tony Jebara (former VP Engineering, Spotify). Trent AI is a partner member of OWASP [EXTERNAL LINK: OWASP Top 10 for LLM Applications] and a startup partner of Carnegie Mellon University’s CyLab Venture Network. Design partners already running the platform include Canopy, Commscentre, ML@Cam, Qbeast, and Weblogic.
What Enterprise Leaders Should Do Now
Trent AI’s emergence is a category signal, not just a funding announcement. Multi-agent security is separating from general AI risk tooling as a distinct infrastructure layer. Enterprise security teams should map this emerging vendor landscape before procurement pressure arrives top-down.
Three immediate actions for enterprise AI and security leads: First, audit which production agentic deployments currently operate without a dedicated runtime security layer. Second, evaluate whether existing SIEM and CSPM tooling was designed to handle autonomous, self-modifying agent behaviour — most was not. Third, track the OWASP Agentic AI security standards as the governance benchmark most enterprise procurement processes will eventually require.
The agentic security multi-agent enterprise category is forming now. The cost of entering it after your first incident is higher than the cost of building the architecture before it.
Source: The Next Web, Trent AI raises $13M to build multi-agent security for a world where AI systems are running themselves
